Inflex, s.r.o. | Legionáøská 1085/8 | 779 00 Olomouc | Czech Republic | info@inflex.cz | tel. 00 420 583 802 583 | www.inflex.cz

Technology

The tool is using its platform to analyze e-mail

communication. The platform provides import

mechanism to collect emails from user email accounts

(Outlook, Thunderbird, IMAP, and Gmail) and also

provides algorithms for pre-processing and analysis of the

collected data.

The e-MailMap tool is using this platform to reveal hidden

characteristics of teams and informal groups. In particular,

answering the questions:

! Which people and which topics are important?

! With whom, when and about what a person or group of

persons communicates?

! Who is important in group communication and how

much, i.e. who is the leader and who is involved only

passively?

! Who, in which group, when and how frequently

communicates about a particular topic?

! How much does communication burden the group?

Moreover, how formalized the communication within

the group is?

! What kind of informal communities exist and what is the

content of their communication?

The following graphical outputs were obtained from one

of the two applications of e-MailMap tool that visualizes

outputs of e-mail communication analysis. The

application was created as part of R&D project LF13030 -

Optimizing of working teams by using the SW for analysis

of the social and professional relations in the company

networks (2013-2015, MSM/LF) – TeamNet for short.

Management support tool

Our softwareis tailored

We build on standards

We develop ourown technologies

RESEARCH

DEVELOPMENT &1/6

Security and Legal Issues

In order to protect users' privacy, only the sender and

recipient of emails and message subjects are processed.

For deeper analysis it is possible to connect the platform

to tools for analyzing the content of emails and their

attachments.

Analytical Background

The platform is based on analyzing the communication

of the group of persons. The analysis is performed

entirely automatically over data that are sent and

received email messages. The analysis uses several

principles, which should be clear before studying the

outputs:

! The basic unit of all analytical measurements is

a conversation. By conversations we mean a set of

email messages originated from a single initial

message while all other messages in the conversation

are directly or indirectly linked to the initial one (as

replies and forwarded messages).

! No messages that are not part of the conversation

affect the analysis.

! The demandingness of each conversation is

calculated. The parameters used are the number of

messages in that conversation, the number of

participating senders, etc.

! The analysis does not work with the people, but with e-

mail accounts (addresses), hereafter denoted as users.

! The share of the conversation is calculated for each

user and each conversation. The parameters used are

the number of messages the user sent or received in

the conversation and in what mode (recipient, in copy).

e-mailMap

Inflex, s.r.o.

Legionáøská 1085/8 | 779 00 Olomouc | Czech Republic | info@inflex.cz | tel. 00 420 583 802 583 | www.inflex.cz

Analytical Outputs

Every output is always related to a selected context for selected period, thus in effect, automatically detected a group of

important users and terms, respectively.

Charts This output provides a list of all users in the group, all terms of the group and users in different roles. These charts

can be ordered according to two criteria:

! Intensity. It expresses in how many messages and how demanding conversations of the group was the user involved or

term included. It is a recalculated share.

! Participation. It expresses, in how many group conversations (regardless of the number of messages) was the user or

term included.

Intensity and participation can be understood as the importance of user/term in group communication in qualitative or

quantitative sense, respectively. A user with high intensity can be interpreted as "communicates a lot" while high

participation as "he knows everything."

2/6

! During the analysis of the content only words from email

subjects, hereafter denoted as terms, are processed.

! For the analysis is used a setting, which may contain one

or more users and one or more terms. Based on this

setting a related group of users and terms is automatically

detected.

! The algorithm for automatic detection of groups is based

on the selection of users and terms from conversations

whose messages contain specified users and terms. These

conversations are hereafter denoted as group

conversations.

! The specified setting can be complemented with the

selection of time interval when the communication took

place. In this case, the same setting of users and terms at

different times can result in different groups. If no interval

is specified, it will be set automatically based on the

setting.

! Selected users and terms together with time interval

represent a group context.

! It may happen that for a group context cannot be found

a non-empty group.

User Roles in Conversation

Depending on how the user behaved in the

conversation, the system recognized six roles. These

roles define user's profile in the group. The first three

roles are active, i.e. users sent, at least, one message

during the conversation. Other three roles are passive,

i.e. users received, at least, one message in

a conversation, but never sent one.

! Initiator. The sender of the first email in the

conversation.

! Solver. The recipient of the first email who sends, at

least, one mail during the conversation.

! Co-solver. The sender of some email in the

conversation.

! Invited. The recipient of the first email in the

conversation.

! Co-invited. The recipient of some email in the

conversation.

! Notified. The recipient of some email in the

conversation (only in CC field).

RESEARCH

DEVELOPMENT &e-mailMapManagement support tool

Activity This output provides a statistical overview of the week and total long-term activity of the group, respectively. In

the week's activity, all sent messages of users are counted. Activity is displayed in the form of the heat map expressing the

color-varying intensity of communication activity, hour after hour in each day of the week. Long-term activity is expressed

as a chart, which shows the total number of sent and received messages in the group over different periods (six months to

a month). The output also provides the total number of messages sent and received in the group and estimated the time

required for their processing.

Groups This output provides an overview of the most important e-mail domains in the analyzed communication. Domain

usually belongs to a cooperating group of people, e.g. from another company. The analysis provides an output with up to

five most intensely communicating groups, including the edges expressing the intensity of their relationship. Other groups

are aggregated into a single group called "others." The number of actively communicating users (senders), the number of

messages sent and, the development of communication intensity in time is provided for each group.

3/6

Inflex, s.r.o.

Legionáøská 1085/8 | 779 00 Olomouc | Czech Republic | info@inflex.cz | tel. 00 420 583 802 583 | www.inflex.cz

RESEARCH

DEVELOPMENT &Management support tool

e-mailMap

Users This output visualizes detected group in the form of a network. Such network can visualize not only the importance

of individual users but also the importance (strength) of relationships between pairs of users, in a transparent way. One of

the algorithms used in the system provides a possibility to uncover the informal internal structure of the group. The

importance of each user is visualized using the so-called tag-cloud.

Topics The network of terms is visualized in this output in a similar fashion as the network of users. Tag-cloud is also used

in the same way to demonstrate the importance of terms. Some group or a tightly connected sub-group of terms can be

understood as a topic of communication of detected group. Each topic is then represented by a term of high importance.

4/6

Inflex, s.r.o.

Legionáøská 1085/8 | 779 00 Olomouc | Czech Republic | info@inflex.cz | tel. 00 420 583 802 583 | www.inflex.cz

RESEARCH

DEVELOPMENT &Management support tool

e-mailMap

Communication Roles This output provides a well-organized view of users arranged according to their importance

within the group. The arrangement can use a qualitative criterion (intensity) or a quantitative criterion (participation). All

roles in which the user is participating in conversations are measured for each user. The share of individual roles in

a conversation of the group can be viewed as a user's profile in that group. One user may have different profiles in various

detected groups. The output also provides statistical information about how much are individual users burdened by

communication.

Communication Flows This output provides an overview of frequently recurring typical patterns that occur in group

conversations. The more frequent and longer the occurring patterns are the more formalized the communication of the

group is. If there are no patterns or if they occur in small quantities in proportion to the number of conversations, it is

a group with less formalized communication. In detected patterns, it is also possible to see who is active in the formalized

communication of the group, and who is not.

5/6

Inflex, s.r.o.

Legionáøská 1085/8 | 779 00 Olomouc | Czech Republic | info@inflex.cz | tel. 00 420 583 802 583 | www.inflex.cz

RESEARCH

DEVELOPMENT &Management support tool

e-mailMap

Analysis of Communication Participants Network properties and the most important people, such as members of

multiple communities, key players, etc., detected from across the whole network of users (or the selected subset) can be

viewed. Individual analysis of selected user from the network is also available. It reveals the structure of his/her

relationships and resulting constraints or benefits.

6/6

Inflex, s.r.o.

Legionáøská 1085/8 | 779 00 Olomouc | Czech Republic | info@inflex.cz | tel. 00 420 583 802 583 | www.inflex.cz

RESEARCH

DEVELOPMENT &Management support tool

e-mailMap